Lwin Moe Naing

Designed, developed, and deployed an internal web tool for tracking and managing manufacturing data, providing visualization for business unit decision-making, and serving as the sole developer from requirements to production release. Built end-to-end pipelines including CI/CD automation, load balancer configuration, VIP setup with multiple servers for high availability, Linux server management, and deployment across multiple environments. Integrated the web tool with SSO, LDAP servers, and other systems to automate previous manual processes. Cross-functional collaboration with AMD MTS, Database Administration team, Server team, and PCBA prototype validation team.

Developed an AI-powered Human Action Recognition system using Python, TensorFlow, and OpenCV, leveraging LSTM and MobileNet models to detect violent behaviors and roadside accidents in real time. Conducted experiments on large-scale video datasets, designed preprocessing pipelines, and optimized model performance for surveillance environments. Presented research findings at the 2025 International Conference on Metaverse and Current Trends in Computing (ICMCTC) and published in IEEE under the title “Real-Time Violence Detection Systems Using MobileNet and LSTM.”